Reconstruction of vancomycin by chemical glycosylation of the pseudoaglycon [20]

Full text of DOI:10.1021/ja982414m

11014
J. Am. Chem. Soc. 1998, 120, 11014-11015
Scheme 1
Reconstruction of Vancomycin by Chemical
Glycosylation of the Pseudoaglycon
Min Ge, Christopher Thompson, and Daniel Kahne*
Department of Chemistry
Princeton UniVersity
Princeton, New Jersey 08544
ReceiVed July 9, 1998
Vancomycin (Figure 1) is a glycopeptide antibiotic that kills
cells by binding to the D-Ala-D-Ala peptide substrate involved in
cross-linking the sugar polymers that comprise the bacterial cell
wall.1 The carbohydrate portion of vancomycin is not directly
involved in binding to D-Ala-D-Ala. However, it has been shown
by scientists at Lilly that N-alkylation of the terminal vancosamine
sugar with a hydrophobic group increases activity against van-
comycin resistant strains dramatically.2,3 Despite the importance
of this carbohydrate in biological activity, no efforts to replace
the vancosamine with a different sugar have been reported.4 In
fact, as far as we know, the chemical glycosylation of vancomycin
at any position has never been achieved.5,6 Vancomycin is a
complex molecule which has a diverse array of functionality and
is sensitive to acid, base, and oxidation.7 Furthermore, it is soluble
only in water and other polar solvents, which are not compatible
with chemical glycosylation reactions. We now report a strategy
for glycosylating the pseudoaglycon of vancomycin. This chem-
istry should permit the synthesis of large numbers of vancomycin
derivatives in which the terminal carbohydrate moiety is varied.
a (a) 1 equiv 1, 2 equiv 2, 2 equiv Tf2O, 4 equiv DTBMP, EtOAc,
-78 °C, 5 min, 92%. (b) 2.5% H2NNH2, MeOH/THF ) 2:1, 3 h, 85%.
(c) 2 equiv 5, 1 equiv Tf2O, 4 equiv DTBMP, Et2O, -78 °C, 0.5 h, 71%.
(d) 2% H2NNH2, MeOH/THF ) 2:1, 8 h, 92%. (e) H2, Pearlman’s
catalyst, MeOH, 1 h, 73%.
recently by Nicolaou.8 The disaccharide contains a â-linkage to
a hindered phenol as well as an R-linkage to a 2-deoxy sugar.
We used 2,6-dimethoxyphenol as a model for the hindered phenol
in vancomycin.8 Our synthesis utilized the sulfoxide glycosylation
reaction9 and proved to be both stereospecific and efficient
(Scheme 1). The equatorial linkage to the 2,6-dimethoxy phenol
was formed in 92% isolated yield by reacting the tin salt of the
phenol10 with a suitably protected sulfoxide 2.11 Stereochemical
control was achieved with neighboring group participation by a
â-ketoester protecting group at C2. This bulky neighboring group
prevents ortho ester formation, and yet it can be readily depro-
tected using mild conditions (i.e., hydrazine).12 Following the
selective removal of the â-ketoester group, the resulting compound
4 was glycosylated on the free C2 hydroxyl with vancosamine
derivative 5.13 The axial isomer was produced exclusively in 71%
yield, and the disaccharide was then deprotected in two steps in
an overall yield of 67%.14 The synthesis of the vancomycin
disaccharide confirmed the utility of the sulfoxide method for
introducing the axial glycosidic linkage to the terminal van-
cosamine using protecting groups that can be removed under mild
conditions compatible with the integrity of the complicated
aglycone. Extending this chemistry to the synthesis of vancomycin
from the pseudoaglycon, however, proved to be considerably more
challenging.
Figure 1.
We initiated our studies to develop chemistry to glycosylate
vancomycin by synthesizing the vancomycin disaccharide. This
structure was first synthesized by Danishefsky in 1992 and more
(8) (a) Dushin, R. G.; Danishefsky, S. J. J. Am. Chem. Soc. 1992, 114,
3471. (b) Nicolaou, K. C.; Mitchell, H. J.; van Delft, F. L.; Rubsam, F.;
Rodriguez, R. M. Angew. Chem. 1998, 110, 1972.
(9) (a) Kahne, D.; Walker, S.; Cheng, Y.; Van Engen, D. J. Am. Chem.
Soc. 1989, 111, 6881. (b) Kim, S.-H.; Augeri, D.; Yang, D.; Kahne, D. J.
Am. Chem. Soc. 1994, 116, 1766. (c) Raghavan, S.; Kahne, D. J. Am. Chem.
Soc. 1993, 115, 1580. For recent studies on the mechanism of the sulfoxide
glycosylation reaction see: (d) Crich, D.; Sun, S. J. Am. Chem. Soc. 1997,
119, 11217.; (e) Gildersleeve, J.; Pascal, R. A.; Kahne, D. J. Am. Chem. Soc.
1998, 120, 5961.
(1) Walsh, C. T.; Fisher, S. L.; Park, I.-S.; Prahalad, M.; Wu, Z. Chem.
Biol. 1996, 3, 21.
(2) (a) Nagarajan, R.; Schabel, A. A.; Occolowitz, J. L.; Counter, F. T.;
Ott, J. L.; Felty-Duckworth, A. M. J. Antibiotics 1989, 42, 63. (b) Nagarajan,
R. J. Antibiotics 1993, 46, 1181. (c) Rodriguez, M. J.; Snyder, N. I.; Zweifel,
M. J.; Wilkie, S. C.; Stack, D. R.; Cooper, R. D.; Nicas, T. I.; Mullen, D. L.;
Butler, T. F.; Thompson, R. C. J. Antibiot. 1998, 51, 560.
(10) Ogawa, T.; Matsui, M. Carbohydr. Res. 1976, 51, C13.
(3) Malabarba, A.; Nicas, T. I.; Thompson, R. C. Med. Res. ReV. 1997,
(11) (a) 3,4,6-Tri-O-Bn glucose11b is converted to 2 in six steps (66%
overall). Acylation (Ac2O, pyridine, CH2Cl2,, 6h, rt), thiophenol displacement
(BF3‚Et2O, PhSH, CH2Cl2, 2h, rt) and acetate hydrolysis (NaOH/MeOH, 3h,
rt) gives 3,4,6-tri-O-Bn glucose phenyl sulfide. Transesterification11c with ethyl-
2-methyl acetoacetate (DMAP, toluene, reflux, 48 h), alkylation (MeI,
potassium-tert-butoxide, THF, 0°, 0.5 h), and oxidation (mCPBA, CH2Cl2,
-60° to -5°, 1 h) gives 2. (b) Betaneli, V. I.; Ovchinnikov, M. V.;
Backinowsky, L. V.; Kochetkov, N. K. Carbohydr. Res. 1982, 107, 285. (c)
Taber, D. F.; Amedio, J. C.; Patel, Y. K. J. Org. Chem. 1985, 50, 3618.
(12) Less-hindered esters often lead to ortho ester side products: (a) Kunz,
H.; Harreus, A. Liebigs Ann. Chem. 1982, 41. (b) Sato, S.; Nunomura, S.;
Nakano, T.; Ito, Y.; Ogawa, T. Tetrahedron Lett. 1988, 33, 4097. (c) Seeberger,
P. H.; Eckhardt, M.; Gutteridge, C. E.; Danishefsky, S. J. J. Am. Chem. Soc.
1997, 119, 10064.
17, 69.
(4) However, vancosamine has been converted to epi-vancosamine.3
(5) The vancomycin aglycon has been enzymatically glycosylated: So-
lenberg, P. J.; Matsushima, P.; Stack, D. R.; Wilkie, S. C.; Thompson, R. C.;
Baltz, R. H. Chem. Biol. 1997, 4, 195.
(6) A great deal of synthetic effort has been focused on the vancomycin
aglycon: (a) Evans, D. A.; Barrow, J. C.; Watson, P. S.; Ratz, A. M.;
Dinsmore, C. J.; Evrard, D. A.; Devries, K. M.; Ellman, J. A.; Rychnovsky,
S. D.; Lacour, J. J. Am. Chem. Soc. 1997, 119, 3419. (b) Nicolaou, K. C.;
Ramanjulu, J. M.; Natarajan, S.; Brase, S.; Li, H.; Boddy, C.; Rubsam, F.
Chem. Commun. 1997, 1899. (c) Boger, D. L.; Beresis, R. T.; Loiseleur, O.;
Wu, J. H.; Castle, S. L. Bioorg. Med. Chem. Lett. 1998, 8, 721. (d) For a
review of work up to 1995, see: Rama Rao, R. V.; Gurjar, M. K.; Reddy, K.
L.; Rao, A. S. Chem. ReV. 1995, 95, 2135.
(7) (a) Nagarajan, R.; Schabel, A. A. J. Chem. Soc., Chem. Commun. 1988,
1306. (b) Harris, C. M.; Kopecka, H.; Harris, T. M. J. Am. Chem. Soc. 1983,
105, 6915. (c) Adamczyk, M.; Grote, J.; Rege, S. Bioorg. Med. Chem. Lett.
1998, 8, 885.
(13) 5 was obtained from vancomycin in 5 steps (48% overall). Protection
of vancomycin amines (Cbz-succinimide, NaHCO3, H2O, dioxane, rt, 4 h)
followed by hydrolysis (1.5 M methanolic HCl, rt, 0.5 h), acylation (Ac2O,
pyridine, CH2Cl2, rt, 2 h), thiophenol displacement (PhSH, BF3‚Et2O, CH2-
Cl2, 0.5 h, rt), and oxidation (mCPBA, CH2Cl2 -78° to -20°, 1 h) gives 5.
10.1021/ja982414m CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/10/1998

Communications to the Editor
J. Am. Chem. Soc., Vol. 120, No. 42, 1998 11015
Scheme 2a
the major compound having a mass consistent with loss of water
from the glycosyl acceptor.17 One of the limitations of the
sulfoxide glycosylation reaction has been that it often does not
proceed cleanly in systems involving amides. We believe that
amide oxygens interfere with glycosylation by reacting with triflic
anhydride and/or with the activated glycosyl donor. It seemed
possible, therefore, that the observed loss of water could result
from conversion of the primary asparagine amide to the nitrile
upon treatment with triflic anhydride.18 A model study using Cbz-
Asn-OBn demonstrated that treatment with triflic anhydride and
base at -78 °C cleanly produced the corresponding nitrile (80%
yield).
We thought it might be possible to suppress the dehydration
of primary amides in the presence of triflic anhydride by
pretreating amide-containing compounds with BF3 and leaving
out the base. In fact, in the Cbz-Asn-OBn model system no nitrile
was produced when BF3 was added prior to triflic anhydride.
Furthermore, a single disaccharide product (9) was isolated in
64% yield (along with 23% of the recovered nucleophile) when
the protected pseudoaglycon was treated with 2 equiv of BF3 prior
to adding Tf2O and donor 11.19 Following a two-step deprotection,
the glycosylated material was found to be identical to vancomycin
1
by H NMR, HPLC, and mass spectral analysis.20
(a) (1) 4 equiv alloc-succinimide, 3 equiv NaHCO3, H2O/dioxane, 3
h; (2) 5 equiv allyl bromide, 2 equiv NaHCO3, DMF, 2 h; (3) 10 equiv
allyl bromide, 5 equiv Cs2CO3, DMF, 6 h, 82%, three steps. (b) 18 equiv
Ac2O, 36 equiv pyridine, 0.2 equiv DMAP, CH2Cl2, 5 h, 91%. (c) 30
equiv BF3‚Et2O, 10 equiv PhSH, CH2Cl2 rt, 0.5 h, 71%. (d) 2 equiv
BF3‚Et2O, 2 equiv Tf2O, CH2Cl2, then 4 equiv 11, Et2O, -78 °C to -20
°C, 1 h, 64% plus 23% recovered 10. (e) 5% hydrazine, allyl alcohol/
MeOH/THF ) 1:1:1, 4 h, 63%. (f) 50 equiv Bu3SnH, 1 equiv
PdCl2(PPh3)3, DMF/AcOH ) 1:1, 10 min, 78%.
The chemistry reported above extends the utility of the
sulfoxide glycosylation reaction to amide-containing systems. We
have been able to achieve the first chemical glycosylation of the
vancomycin pseudoaglycon, possibly the most complex system
that has yet been glycosylated using a chemical approach. Our
next goal is to develop chemistry to construct the glycosidic
linkage to the phenol on the vancomycin aglycon, a considerable
challenge in its own right. Accomplishing this objective is
essential to completing the total synthesis of vancomycin.6 In the
meantime, the present work permits further exploration of the
role of the carbohydrate portion of vancomycin in antibiotic
activity, particularly in conferring activity against resistant
microorganisms.
Glycosylation required a suitably protected vancomycin pseudo-
aglycon. This compound had to have protecting groups that would
confer solubility in nonpolar organic solvents at -78 °C. These
protecting groups had to be removable under conditions that would
not damage the final product. After considerable experimentation,
we established an efficient degradation route to 10 (Scheme 2)
by sequentially protecting the various functional groups in
vancomycin according to their reactivities, with the two amines
acylated first with alloc-succinimide followed by alkylation of
the carboxylic acid and then the three free phenols with allyl
bromide. Allyl-derived protecting groups were selected for all of
these functional groups because they can be removed cleanly in
one step with PdCl2(PPh3)3-Bu3SnH.15 Finally, the primary and
five secondary hydroxyls were protected as acetates. The glyco-
sidic linkage to the vancosamine was then selectively cleaved
with BF3-thiophenol to produce 10.
Acknowledgment. This work was supported by the National Institutes
of Health and Interneuron Pharmaceuticals, and Merck Research
Laboratories.
JA982414M
(17) The IR specrum of the major product shows a weak stretch at 2248
cm-1
.
(18) Dehydration of amides upon treatment with triflic anhydride and
pyridine has been reported: Charette, A. B.; Chua, P. Synlett 1998, 163.
(19) To produce 9, 10 (22.4 mg, 0.0127 mmol) was azeotroped 3× with
toluene, dissolved in 1 mL of CH2Cl2 and cooled to -78 °C. BF3Et2O (2.3
µL, 0.0191 mmol) was added followed by triflic anhydride (4.3 µL, 0.0255
mmol). Sulfoxide 11 (20 mg, 0.0509 mmol) in 0.5 mL of CH2Cl2 was then
added dropwise over 1 min, and the reaction was allowed to warm to -25 °C
over 1.5 h. After the reaction was quenched with a 1:1 solution of MeOH/
DIEA (100 µL), the solvent was removed under vacuum, and the residue was
chromatographed over silica gel (10 mm × 5 cm column; 20 mL of 100%
CHCl3; 20 mL of 2.5% MeOH/CHCl3; 20 mL of 5% MeOH/CHCl3). Impure
product (25 mg) was obtained and further purified by reverse-phase HPLC
(PHENOMENEX LUNA C18 column, 21 × 25 mm, 5 µM particle size) using
a linear gradient (80% CH3CN/0.1% CH3COOH/19.9% H2O to 99.9% CH3-
CN/0.1% CH3COOH over 30 min). 9 (16.5 mg, 64%) was obtained along
with 5 mg of recovered 10 (23%). Rf ) 0.3 (5% MeOH/CHCl3). HRMS
(FAB): calcd for C98H111N9NaO34Cl2 [M + Na+] 2050.6508, found 2050.6458.
(20) Synthetic vancomycin coeluted with authentic vancomycin on reverse-
phase analytic HPLC using a PHENOMENEX PRODIGY 5 µm ODS(3) 100
Å column (250 × 4.6 mm), eluting with 0.1% TFA in water for 2 min followed
by a linear gradient of 0.1% TFA in water to 20% CH3CN/0.1% TFA in water
over 28 min. Vancomycin elutes at 27.9 min. Mass: [M + H]+, 1449.3, [M
- V + H] +, 1305.2, [M - V - G + H] +, 1144.3.
Initial efforts to glycosylate the protected pseudoaglycon 10
using sulfoxide 1116 and triflic anhydride in the presence of base
(DTBMP) yielded an unencouraging mixture of products, with
(14) Spectral data for 7: 1H NMR (CD3OD, 500 MHz) δ 7.03 (t, J ) 8.2
Hz, 1 H), 6.68 (d, J ) 9.2 Hz, 2 H), 5.39 (d, J ) 4.0 Hz, VH-1, 1 H), 5.16 (J
) 7.6 Hz, GH-1, 1 H), 4.55-4.52 (m, VH-5, 1 H), 3.81 (s, OMe, 6 H), 3.68-
3.60 (m, GH-2, GH-6′, GH-6, 3 H), 3.51 (t, J ) 9.1 Hz, GH-3, 1 H), 3.42 (t, J
) 9.5 Hz, GH-4, 1 H), 3.20 (s, VH-4, 1 H), 3.14-3.11 (m, GH-5, 1 H), 2.04-
1.93 (m, VH-2, VH-2′, 2 H), 1.65 (s, CH3, 3 H), 1.07 (d, J ) 6.7 Hz, CH3, 3
H); 13C NMR (CD3OD, 500 MHz) δ 154.90, 134.00, 125.66, 107.30, 101.95,
98.59, 79.50, 79.27, 78.17, 73.29, 71.65, 64.97, 62.62, 56.91, 55.60, 35.14,
23.68, 17.08. HRMS (FAB): calcd for C21H33NNaO10 [M + Na+] 482.1991,
found 482.2002.
(15) Dangles, O.; Guibe, F.; Balavoine, G.; Lavielle, S.; Marquet, A. J.
Org. Chem. 1987, 52, 4984.
(16) 11 was prepared in the same way as 5, substituting alloc-succinimide
for Cbz-succinimide.